CN101277366A - Image processing device, image processing method - Google Patents

Image processing device, image processing method Download PDF

Info

Publication number
CN101277366A
CN101277366A CNA2008100878720A CN200810087872A CN101277366A CN 101277366 A CN101277366 A CN 101277366A CN A2008100878720 A CNA2008100878720 A CN A2008100878720A CN 200810087872 A CN200810087872 A CN 200810087872A CN 101277366 A CN101277366 A CN 101277366A
Authority
CN
China
Prior art keywords
data
attribute data
attribute
pixel value
image
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CNA2008100878720A
Other languages
Chinese (zh)
Inventor
斋鹿尚史
八幡洋一郎
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sharp Corp
Original Assignee
Sharp Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sharp Corp filed Critical Sharp Corp
Publication of CN101277366A publication Critical patent/CN101277366A/en
Pending legal-status Critical Current

Links

Images

Abstract

The image processing device of the present invention includes: an image acquiring section ( 2 ) acquiring image data; an image data generating section ( 32 ) generating attribute data including each attribute value indicative of an attribute to which each of a plurality of pixels constituting an image indicated by the image data acquired by the image acquiring section ( 2 ) belongs among predetermined attributes; and image data reducing section ( 34 ) reducing the attribute data by converting, to one attribute value, at least two attribute values included in the attribute data generated by the attribute data generating section, according to a priority given to the each attribute value. Therefore, it becomes possible to realize an image processing device that can reduce the attribute data while keeping important information included in the attribute data, and further improve encoding efficiency of the attribute data.

Description

Image processing apparatus and image processing method
Technical field
The present invention relates to use the attribute data of attribute of the pixel of expression composing images to carry out image processing apparatus, the image processing method of image processing and write down the recording medium of image processing program.
Background technology
In recent years, image processing apparatus with function that image is stored as electronic data such as digital copier is popularized.Think with the hard disk to be that the high capacity of storage device and the low price of representative influences popularizing of this image processing apparatus.
Image processing apparatus uses the attribute data of the attribute of the pixel of representing composing images to carry out image processing.One example of this image processing apparatus is disclosed the spy and opens in the 2001-211316 communique (August 3 calendar year 2001 is open).Open in the image processing apparatus of 2001-211316 communique record the spy, will be stored in from the view data of scanner input with by the attribute data (flag data) that this view data generates in the auxilary unit such as hard disk.When print image, from hard disk, take out this view data and attribute data.
Above-mentioned attribute data is normally represented the data of each attributes of images (literal, photo etc.), is used for image processing when print image etc. constantly.Below, the attribute of so-called pixel is meant that this pixel of expression belongs to the information that each pixel is provided in which zone (for example character area, photo zone, background area etc.) in the image.
As the example of image processing, enumerate when printing and reduce under the situation of tonal gradation, switch the processing of dither matrix (dither matrix) etc. according to attribute.In addition, if in reprocessing Useful Information, then also the presentation graphs pixel can be belonged to data outside the information in which zone as attribute data.
The data that to carry out each regional attribute of comprising in each pixel of presentation video of this useful for image processing or the image below show as " attribute data ".
Here, use Figure 28 to illustrate that what kind of data attribute data is.This figure is the figure of an example of expression input picture 101.As shown in the drawing, establishing provides the input picture 101 that is made of photo zone, character area and background area.The figure that amplifies in order to be illustrated as the subregion 102 of the part of input picture 101 is Figure 29.
Subregion 102 is made of character area 201 (being made of the pixel that constitutes " UK " each literal), photo zone 202 and background area 203.Figure 30 is the figure of the expression attribute data corresponding with subregion 102.Attribute data itself also has the structure as view data, and each pixel of attribute data is physically corresponding with the pixel of input image data.
As shown in figure 30, the zone that embeds with identical patterns has same pixel value (property value).For example, the pixel value of establishing the pixel that belongs to the zone of representing with the horizontal line pattern 301 is 1, and the pixel value that belongs to the pixel in the zone of representing with oblique line 302 is 2, and the pixel value that belongs to the pixel in the zone of representing with point 303 is 0.If the value of the attribute data integral body corresponding with input pixel 101 all shows with situation shown in Figure 30, then attribute data is represented with 3 values.
In addition, here, being illustrated as attribute data is identical resolution with input image data, but the resolution of the resolution of input image data and attribute data there is no need consistent.As long as it is just passable to judge which classification is each pixel that comprises in the image that input image data represents belong to by attribute data.
In order to store the view data of the many images of number of pages, need to increase storage area, but under the situation that can not increase storage area, need reduce the desired volume of each page.
For the storage area of each page of reducing image, think the resolution or the pixel count that will reduce image, but if simple extracted data, then the information dropout of image is bad.
This situation is for view data itself, and is all identical for attribute data, but under the situation of view data, the method for dwindling amplification at not reducing picture quality has a lot of research and development before this.The one example is documented in " newly organized graphical analysis handbook " (Tokyo University publishing house, on September 10th, 2004,1350-1373 page or leaf).
This method is thought basically because the pixel value of view data is not given implication especially, so feasible.Relative therewith, under the situation of attribute data, because the value (property value) of each pixel itself has implication, so the situation difference.
For example, establishing the implication of carrying out the pixel value of attribute data as shown in figure 31 like that gives.This figure is the table of corresponding relation of the pixel value of the attribute of pixel of presentation video data and this attribute of expression.At this moment, when dwindling attribute data, have two pixels of pixel value " 3 " (site) and pixel value " 5 " (photo) respectively, so, then become the pixel value " 4 " of expression colored text if get this mean value because of existing.Therefore, the processing of dwindling attribute data by the mean value of calculating pixel value does not have implication.
Therefore, the spy opens and discloses in the 2001-211316 communique when the dwindling of each attribute data (flag data), prevent the method that specific value (for example 1) is lost and around concerned pixel, count the frequency of occurrences of particular value (for example 1) by contiguous neighbor list (near list neighbor) method of dual-purpose and logic OR method, export the method for 1 grade at this counting under for certain above situation.
But the spy opens in the method for 2001-211316 communique record, is difficult to want in the reserved property data the reliable pixel value that keeps.For example, in the method based on the frequency of occurrences of pixel value, even wish the pixel value that keeps, but under the low situation of the frequency of occurrences of this pixel value, the possibility that keeps this pixel value is also low.
With logic-based or mode be applicable under the situation of the multivalue image that constitutes by a plurality of bit planes, want the reliable pixel value that keeps to need all positions to be upgraded to 1.If not like this,, keep so can not guarantee the pixel value of original preservation owing to may not be certain to constitute the position reservation of this value by bit plane.Therefore, logic-based or mode in, change the pixel value want to keep and be not easy, provide high priority also very difficult to a plurality of pixel values.
Another problem is attribute data to be decomposed under the situation of bit plane, might be because of the deciding means of pixel value, and code efficiency reduces.
The following describes this situation.As shown in figure 31, the pixel value of attribute data is being given under the situation of implication, if input then is decomposed under the situation of bit plane at the pixel value that will represent these two attributes by the original copy that background and black literal lines constitute, three bit planes are different fully.Therefore, this two kinds of pixel values account for a greater part of possibility height when the text original copy, but are same value or the bit plane approaching with this value owing to do not generate whole, separate mapped coding efficient so can not improve the zone.
Therefore, Figure 32 represent to switch with black literal lines and colored text lines corresponding pixel value after table.At this moment, do not become problem in this case,, although colored text lines and background exist simultaneously with the very high frequency of occurrences in the original copy of colour, but produced problem same as described above in this case, so can not improve the code efficiency of attribute data.
Special generation method of being willing to put down in writing in the flat 4-134570 communique (on May 8th, 1992 is open) downscaled images of faithful reappearance original image.But, the not record relevant in this communique with dwindling of attribute data.
Summary of the invention
The purpose of this invention is to provide and a kind ofly both can keep the important information that comprises in the attribute data, can dwindle the image processing apparatus of attribute data again, a kind of image processing apparatus that can improve the code efficiency of attribute data further is provided.
To achieve these goals, the image processing apparatus that the present invention relates to is characterized in that, comprising: obtain the unit, it obtains view data; The attribute data generation unit, its generation comprises the attribute data of property value, and this property value is represented to constitute the above-mentioned a plurality of pixels that obtain the represented image of the view data that obtains the unit and is belonged in the attribute that is predetermined which respectively; Attribute data dwindles the unit, and it is transformed to a property value according to the priority that above-mentioned each property value is given by at least two property values that comprise in the attribute data that above-mentioned attribute data generation unit is generated, and dwindles this attribute data.
The image processing method that the present invention relates to is characterized in that, comprising: obtain step, obtain view data; Attribute data generates step, generates the attribute data that comprises property value, and this property value is represented to constitute the above-mentioned a plurality of pixels that obtain the represented image of the view data that obtains in the step and belonged in the attribute that is predetermined which respectively; The attribute data reduction process according to the priority that above-mentioned each property value is given, is transformed to a property value by above-mentioned attribute data being generated at least two property values that comprise in the attribute data that generates in the step, dwindles above-mentioned attribute data.
According to said structure, the attribute data generation unit generates the attribute data comprise property value, and this property value represents which attribute is a plurality of pixels that formation obtains the represented image of the view data that obtains the unit belong to respectively.Attribute data dwindles the unit according to the priority that each property value is given, and is transformed to a property value by at least two property values that comprise in the attribute data that the attribute data generation unit is generated and dwindles attribute data.
Therefore, the high property value of priority can be reliably kept, attribute data can be dwindled simultaneously.
The recording medium that has write down the embodied on computer readable of image processing program of the present invention is characterised in that, makes computer carry out following the processing: to obtain processing, obtain view data; Attribute data generate to be handled, and generates the attribute data that comprises property value, and this property value is represented to constitute the above-mentioned a plurality of pixels that obtain the image of the pictorial data representation of obtaining in the step and belonged in the attribute that is predetermined which respectively; Attribute data dwindles processing, according to the priority that above-mentioned each property value is given, is transformed to a property value by generating at least two property values that comprise in the attribute data of handling generation by above-mentioned attribute data, dwindles above-mentioned attribute data.
Therefore, can make computer carry out the reliable high property value of priority that keeps, dwindle the processing of attribute data simultaneously.
Other purposes of the present invention, feature and advantage can fully be understood by the record shown in following.Become clear in the explanation of advantage of the present invention below with reference to the accompanying drawings.
Description of drawings
Fig. 1 is the functional block diagram of structure of the attribute data generating unit that has of image processing apparatus of expression one embodiment of the present invention;
Fig. 2 is the functional block diagram of structure of the image processing apparatus of expression one embodiment of the present invention;
Fig. 3 is the flow chart of an example of the handling process in the above-mentioned image processing apparatus of expression;
Fig. 4 is the flow chart of an example of the handling process in the representation attribute data generating unit;
Fig. 5 is the figure of an example of remarked pixel value priority list;
Fig. 6 is the flow chart that the presentation video data are dwindled an example of the handling process in the portion;
Fig. 7 is the figure of an example of the block diagram of representation attribute data;
Fig. 8 is the figure of an example of remarked pixel value transform table;
Fig. 9 is the table of the pixel value before and after the conversion carried out of presentation video data shaping portion;
Figure 10 is the flow chart of an example of the handling process in the presentation video data shaping portion;
Figure 11 is the functional block diagram of the structure of representation attribute digital coding portion;
Figure 12 is the flow chart of an example of the handling process in the representation attribute digital coding portion;
Figure 13 is the functional block diagram of the structure of representation attribute data decoding part;
Figure 14 is the flow chart of an example of the handling process in the representation attribute data decoding part;
Figure 15 is the functional block diagram of the structure of representation attribute adjustment of data portion;
Figure 16 is the figure of an example of the amplification method of the attribute data in the presentation video data enlarging section;
Figure 17 is the flow chart of an example of the handling process in the representation attribute adjustment of data portion;
Figure 18 is the flow chart of an example of the pixel value map table that uses of presentation video data recovery section;
Figure 19 is the flow chart of an example of the handling process in the presentation video data recovery section;
Figure 20 is the functional block diagram of structure of the image processing apparatus of expression another embodiment of the present invention;
Figure 21 is the functional block diagram of the structure of the attribute data generating unit that has of the above-mentioned image processing apparatus of expression;
Figure 22 is the functional block diagram of the structure of the view data shaping portion that has of the above-mentioned image processing apparatus of expression;
Figure 23 is the flow chart of an example of the handling process in the presentation video data shaping portion;
Figure 24 is the functional block diagram of the structure of remarked pixel value transform table generating unit;
Figure 25 (a) is the figure of expression by a histogrammic example of histogram generating unit generation;
Figure 25 (b) is the figure of expression by an example of the sequence list of sequence list generating unit generation;
Figure 26 is the figure of an example of order of representation pixel value map table;
Figure 27 is the flow chart of an example of the handling process in the remarked pixel value transform table generating unit;
Figure 28 is the figure of an example of expression input picture;
Figure 29 is the figure of expression as the subregion of the part of above-mentioned input picture;
Figure 30 is the figure of the expression attribute data corresponding with above-mentioned subregion;
Figure 31 is the table of corresponding relation of the pixel value of the attribute of pixel of presentation video data and this attribute of expression;
Figure 32 is another table of corresponding relation of the pixel value of the attribute of pixel of presentation video data and this attribute of expression.
Embodiment
[execution mode 1]
If an embodiment of the invention are described, then as follows according to Fig. 1～Figure 19.
(structure of image processing apparatus 1)
Fig. 2 is the functional block diagram of structure of the image processing apparatus 1 of expression present embodiment.As shown in the drawing, image processing apparatus 1 comprises that view data obtaining section (obtaining the unit) 2, attribute data generating unit (attribute data generation unit, attribute data dwindle the unit) 3, attribute data encoding section 4, attribute data storage part 5, attribute data lsb decoder 6, attribute data correction unit 7, view data dwindle portion 8, coded image data portion 9, image data storage portion 10, image data decoding portion 11, view data enlarging section 12 and view data efferent 13.
This image processing apparatus 1 can link to each other with image-data output device 15 with view data input unit 14 communicatedly.View data input unit 14 is devices of input image data.Can enumerate for example scanner as view data input unit 14.
View data obtaining section 2 obtains view data from view data input unit 14, and this view data is outputed to attribute data generating unit 3 and view data is dwindled portion 8.
The view data that attribute data generating unit 3 obtains according to view data obtaining section 2, generation comprises the attribute data of pixel value (property value), and which in the attribute that is predetermined be a plurality of pixels of the image that this view data of this pixel value (property value) expression formation is represented belong to respectively.The details of the generation method of this attribute data is described in the back.
The attribute data that 4 pairs of attribute data generating units of attribute data encoding section 3 generate is encoded, and the attribute data after will encoding stores in the attribute data storage part 5.
Attribute data in 6 pairs of attribute data storage parts 5 of attribute data lsb decoder behind the coding of storage is decoded, and decoded attribute data is stored in the attribute data storage part 5 once more.
Attribute data correction unit 7 is proofreaied and correct by attribute data lsb decoder 6 decoded attribute datas.The bearing calibration of this attribute data is described in the back.
View data is dwindled portion 8 and is dwindled the view data identical with the view data that is provided for attribute data generating unit 3.Dwindle dwindling method and can using suitable existing method of view data in the portion 8 as view data.In addition, in the present embodiment, view data is dwindled portion 8 view data is reduced into 1/2 respectively in longitudinal and transverse direction.
9 pairs of view data of dwindling after portion 8 dwindles by view data of coded image data portion are encoded, and encoded image data is stored in the image data storage portion 10.Can use suitable existing method as the coding method in the coded image data portion 9.For example, can enumerate JPEG (Joint Photographic Experts Group JPEG (joint photographic experts group)) mode and PNG (mobile network's figure Portable Network Graphic) mode as representative, but be not limited to this.
11 pairs in image data decoding portion decodes by coded image data portion 9 encoded image data, and with decoded image data storage in image data storage portion 10.The mode of the decoding of using in the image data decoding portion 11 is corresponding to the coded system of using in the coded image data portion 9.
View data enlarging section 12 enlarged image data decoding part, 11 decoded view data.In the view data enlarging section 12, for example can use aforesaid suitable image magnification method.In the present embodiment, carry out 1/2 the concord mutually of dwindling respectively, in longitudinal and transverse direction, be enlarged into 2 times respectively with longitudinal and transverse direction.Certainly, this is consistent with the resolution of the image of final output because of the resolution of the image that makes 14 inputs of view data input unit, but inessential as if this, then magnification ratio can be a numerical value in addition, can omit the operation that image amplifies according to purpose.
View data efferent 13 is obtained by the attribute data after 7 corrections of attribute data correction unit with by the view data after 12 amplifications of view data enlarging section, and these data are outputed to image-data output device 15.
Image-data output device 15 will be exported as image from the view data of view data efferent 13 outputs according to the attribute data from 13 outputs of view data efferent.The representative of image-data output device 15 is laser printers.Use for the attribute data here, opening the 2001-211316 communique as the spy puts down in writing, can change the method that loses lustre in when output by the value of attribute data, or the coefficient when changing the color space transformation that is used for from RGB to CMYK, or change the degree of definition.
(flow process of the processing in the image processing apparatus 1)
Fig. 3 is the flow chart of an example of the flow process of the processing in the presentation video processing unit 1.
At first, with the input image data input, view data input unit 14 outputs to view data obtaining section 2 (step S1) with this view data from view data input unit 14.
View data obtaining section 2 obtains view data from view data input unit 14, and this view data is outputed to attribute data generating unit 3 and view data is dwindled portion 8.
Attribute data generating unit 3 generates attribute data (step S2) from the view data of view data obtaining section 2 outputs, and the attribute data that is generated is outputed to attribute data encoding section 4.
The attribute data that 4 pairs of attribute data generating units of attribute data encoding section 3 generate is encoded (step S3), and the attribute data after will encoding stores in the attribute data storage part 5.
On the other hand, view data is dwindled portion 8 receiving after the view data of view data obtaining section 2 outputs, dwindles this view data (step S4), and the view data after this is dwindled outputs in the coded image data portion 9.
After coded image data portion 9 receives and dwindles view data after portion 8 dwindles by view data, this view data of encoding (step S5), and encoded image data stored in the image data storage portion 10.
After turning back to step S1 from step S5, the attribute data corresponding with multipage stored in the attribute data storage part 5, owing to the action that can carry out the attribute data corresponding with multipage stored in the attribute data storage part 10 is very important from the viewpoint of reality use, but inessential for the understanding of invention, so with dashed lines is simply represented and is kept in flow chart shown in Figure 3.
Viewpoint from the reality use, by improving the code efficiency (reducing the data volume of the attribute data under the encoding state) of attribute data, can reduce the capacity of the required attribute data storage part 5 of memory attribute data, image processing apparatus 1 price is lower very important.In addition, reduce data volume and also bring on general the data of device inside to transmit to become at a high speed, by shortening user's stand-by period, thereby also bring the convenience of use to improve such advantage.
Obtain in the attribute data lsb decoder 6 dependency data store 5 by the attribute data behind attribute data encoding section 4 codings, and this attribute data (step S6) of decoding.Attribute data lsb decoder 6 stores decoded attribute data in the attribute data storage part 5 into once more.
Obtain by attribute data lsb decoder 6 decoded attribute datas in the attribute data correction unit 7 dependency data store 5, and proofread and correct this attribute data (step S7).Attribute data after attribute data correction unit 7 will be proofreaied and correct outputs to view data efferent 13.
On the other hand, image data decoding portion 11 obtains 9 encoded image data by coded image data portion from image data storage portion 10, and this view data (step S8) of decoding.Image data decoding portion 11 with decoded image data storage in image data storage portion 10.
Then, image data decoding portion 11 decoded view data are obtained in view data enlarging section 12 from image data storage portion 10, and amplify this view data (step S9).View data after view data enlarging section 12 will be amplified outputs to view data efferent 13.
View data efferent 13 is obtained by the attribute data after 7 corrections of attribute data correction unit with by the view data after 12 amplifications of view data enlarging section, and these data are outputed to (step S10) in the image-data output device 15.
The attribute data that image-data output device 15 bases are exported from view data efferent 13, and will export as image from the view data of view data efferent 13 outputs.
(structure of attribute data generating unit 3)
Fig. 1 is the functional block diagram of the structure of representation attribute data generating unit 3.As shown in the drawing, attribute data generating unit 3 has controller 31, view data generating unit (attribute data generation unit) 32, image data buffer 33, view data and dwindles portion's (attribute data dwindles the unit) 34 and view data shaping portion 35.Each inscape can connect each other communicatedly through data/address bus, and links to each other with the inscape of outside.
Each inscape in the controller 31 controlled attribute data generating units 3.
View data generating unit 32 generates the attribute data that comprises pixel value (property value), and be stored in the image data buffer 33, wherein which in the predetermined attribute be a plurality of pixels of the represented image of the view data that obtains of this pixel value (property value) expression composing images data obtaining section 2 belong to respectively.As the attribute data generation method in the view data generating unit 32, for example can use the whole bag of tricks that separates as disclosed image field in the patent documentation 1.
In the following description, view data generating unit 32 identification backgrounds, site (halftone dot), colored text, photo and black literal, and, generate attribute data as the pixel of the pixel of view data and attribute data image one to one.That is, the pixel value of the pixel of attribute data is the property value of the attribute of the pixel that comprises in the image of expression by pictorial data representation.The pixel value that comprises in the attribute data that view data generating unit 32 generates for example is the value shown in the left column of table shown in Figure 31.
View data is dwindled the attribute data of view data generating unit 32 generations of storage in portion's 34 down scaling image data buffers 33, and the attribute data after will dwindling is stored in the image data buffer 33 once more.In more detail, view data is dwindled portion 34 according to the priority that each property value is given in advance, at least two pixel values (property value) by comprising in the attribute data of replacing 32 generations of view data generating unit with the higher pixel value (property value) of the priority that comprises in this attribute data dwindle attribute data.In other words, view data is dwindled portion 34 above-mentioned at least two property values is transformed to a property value.The back is described this view data and is dwindled the details of dwindling method in the portion 34.
35 shapings of view data shaping portion are dwindled attribute data after portion 34 dwindles by view data, and via controller 31 outputs to the attribute data after the shaping in the attribute data encoding section 4.In more detail, the pixel value of view data shaping portion 35 conversion attribute datas makes the little possibility of Entropy Changes of each bit plane when attribute data is decomposed into bit plane improve, that is, code efficiency is improved.
(flow process of the processing in the attribute data generating unit 3)
Fig. 4 is the flow chart of an example of the flow process of the processing in the representation attribute data generating unit 3.In addition, even under the situation of other inscapes of 31 control of controller, do not put down in writing this situation especially yet, and be recited as the action of this inscape.
At first, controller 31 is after obtaining view data from view data obtaining section 2, and order view data generating unit 32 generates attribute data.
After accepting this order, view data generating unit 32 generates attribute data from the view data of view data obtaining section 2 output, and stores (step S11) in the image data buffer 33 into.
Afterwards, view data is dwindled portion 34 and obtain the attribute data that view data generating unit 32 generates from image data buffer 33, and after dwindling this attribute data (step S12), the attribute data after will dwindling stores in the image data buffer 33 once more.
Afterwards, view data shaping portion 35 obtains view data and dwindles attribute data after portion 34 dwindles from image data buffer 33, behind this attribute data of shaping (step S13), the attribute data via controller 31 after the shaping is outputed to attribute data encoding section 4 (step S14).
(view data is dwindled the method for dwindling of attribute data in the portion 34)
View data is dwindled portion 34 and is had pixel value priority list 34a, with reference to this pixel value priority list 34a, and the attribute data that comes down scaling image data generating unit 32 to generate.For example, view data is dwindled the property value of the attribute data of portion 34 by replacing the corresponding pixel of a plurality of pixels of comprising in the presumptive area in the image of obtaining with view data obtaining section 2 (below be called piece) with a highest property value of the property value medium priority of a plurality of pixels that comprise in this piece, and dwindles attribute data.Several property values are reduced into a property value as described later, and the minification that depend on attribute data decides.
Pixel value priority list 34a is to be keyword (key) with the appropriate value as the pixel value that comprises in the attribute data, and the attribute of the pixel that comprises in the image that pixel data is represented is got the table that preset priority is used.That is, pixel value priority list 34a is the attribute of remarked pixel and the table of the corresponding relation of the value of the priority of this attribute of expression.Therefore, if pixel value priority list 34a is the table that the priority of attribute extracted out in keyword with 0～7 such value for example, just very abundant.
In the following description, priority is zero or positive integer, and be interpreted as: the value of priority is big more, and priority is high more.
The table that schematically shows among Fig. 5 is illustrated as pixel value priority list 34a.Fig. 5 is the figure of the example of remarked pixel value priority list 34a.Pixel value shown in the pixel value priority list 34 among this figure is equivalent to pixel value shown in Figure 31.As if trying and Figure 31 comparison, then provide the highest priority " 3 " to pixel value " 7 " (black literal lines) as can be seen, then, provide priority " 2 " to pixel value " 4 " (colored text lines), provide priority " 1 " to pixel value " 5 " (photo) and pixel value " 3 " (site), provide priority " 0 " such priority pixel value " 0 " (background).That is, the background priority as pixel value " 0 " is minimum.
In addition, to not distributing pixel value (being the part of " N/A " among Figure 31) input " 0 " such priority of corresponding attribute, still importing what kind of priority in this case can be not influential to moving in pixel value priority list 34a shown in Figure 5.
Decision is that pixel value is big more among the pixel value priority list 34a, the form that priority is high more, and this is in order to understand explanation easily, on the contrary, if determine that priority is high more for pixel value is more little, then also not change within the scope of the invention.Importantly the pixel value to each pixel of attribute data decides priority.
Here,, use the method that shows with the such form of pixel value priority list 34a, be embedded into medium other the demonstration form of program but also can use as conditional statement as the method for the priority of remarked pixel value.The advantage that the changes such as replacement that can carry out priority wherein, are arranged with the priority that shows pixel value simple and reliablely.
In the said structure, the content of pixel value priority list 34a can be predetermined, even but perhaps these make up this content of adaptive determining also no problem for the interior perhaps pixel data of dependency data interior.
Promptly, image processing apparatus 1 has priority generating unit (the priority generation unit does not illustrate), its according to the content of the interior perhaps attribute data of view data one of at least, the priority that decides each property value to the pixel of view data to provide, view data is dwindled portion 34 also can be by the priority of basis by the decision of priority generating unit, be transformed to a property value by at least two property values that comprise in the attribute data that view data generating unit 32 is generated, and dwindle this attribute data.
For example, the priority generating unit also can comprise in the attribute data that is generated by view data generating unit 32 has the certain threshold value of ratio with the pixel of black literal corresponding pixel value " 7 " when big, to providing higher priority with background corresponding pixel value " 0 ".According to this structure, can prevent to have the major part of occupying the attribute data after dwindling with the pixel of black literal corresponding pixel value.
Perhaps, the priority generating unit also can be by the distribution of reference from the pixel value of the view data of view data obtaining section 2 outputs, in attribute data, investigate and have and distribute, and distribute according to this color of pixel and to decide priority corresponding to the corresponding color of pixel of the pixel of the pixel value of background.
For example, when the shared ratio of pixel that can bright in background (brightness height) is big and the shared ratio of the pixel of dark (brightness is low) when big, change the priority of the pixel corresponding by attribute data with the black literal.Consideration occupies under the situation of background in dark pixel, because be judged as the demand height of emphasizing the black literal, thus the priority of black literal improved, but occupy under the situation of background in bright pixel, it is not high to be judged as the demand of emphasizing the black literal, and does not improve the priority of black literal.Certain this situation only is an example, importantly can come adaptive determining priority according to the content of view data and attribute data.
(view data is dwindled the flow process of the processing in the portion 34)
Fig. 6 is the flow chart that the presentation video data are dwindled an example of the handling process in the portion 34.At first, view data is dwindled portion 34 comes storage in the scan image data buffer 33 by each piece attribute data (step S21).
Fig. 7 is the figure of an example of the piece of representation attribute data.Here said (predetermined zone) decides by the minification of attribute data.If attribute data is reduced into 1/2 respectively in length and breadth, then become zone shown in Figure 7 by horizontal 2 pixels * vertical 2 pixels constitute.That is, attribute data is being reduced into respectively under the situation of 1/N in longitudinal direction and transverse direction, one side piece is N pixel.Below, be assumed to and dwindle attribute data with this minification and be illustrated.
Among Fig. 7, become simple in order to make explanation, the attribute data of storage in the image data buffer 33 is illustrated as the image of vertical 4 pixel * 4 pixels of horizontal stroke, each zone of 2 pixels of 2 pixels of horizontal stroke of being surrounded by black surround * indulge is a piece.Arrow 16 presentation video data are dwindled the order that portion 34 is scanned.That is, view data is dwindled portion 34 and is extracted the pixel value of a plurality of pixels that comprise in the piece out with predetermined order.
View data is dwindled in the pixel value that portion 34 comprises in selecting above-mentioned, the pixel value (step S22) of expression greatest priority.
Here said greatest priority represents that by the pixel value priority list 34a that has illustrated for example, if the pixel value of the pixel of attribute data is 7, then its priority is 3.
For there being a plurality of situations,, determine to making it the preceding pixel that occurs preferential with scanning sequency shown in Figure 7 with pixel of same priority.Certainly, as long as purpose is arranged, also can be other determining methods.
Then, view data is dwindled the pixel value that portion 34 will work as the pixel that comprises in the piece in the pre-treatment and is replaced with the pixel value of selecting in the step 22 (step S23).That is, owing to replace horizontal 2 pixels * vertical 2 pixels with a pixel, so dwindled attribute data.
Like this, carry out dwindling of attribute data, dwindling when handling at every turn, but all in the reserved property data, want the pixel value (pixel value that priority is high) that keeps by priority according to pixel.This method of dwindling has been given pixel value itself under the situation of implication as attribute data, compares with the situation of dwindling common image, and importance is especially high.
Especially,, dwindle under the situation of the attribute data that uses simultaneously with view data as the present invention, different with the situation of dwindling attribute data separately, attribute data dwindle the output result that the result influences view data.Therefore, think that dwindling attribute data has special effect according to the priority that offers each pixel value.
Then, view data is dwindled portion 34 and is judged whether to have handled all pieces (step S24), if do not handle all pieces (being not among the step S24), processing enters into next piece (step S25), and get back to S21, if handled all pieces (among the step S24 for being), then finish a series of processing.
(the attribute data shaping methods in the view data shaping portion 35)
View data shaping portion 35 has pixel value map table 35a.This pixel value map table 35a is that the appropriate value with as the pixel value that comprises in the attribute data that dwindles by view data after portion 34 dwindles that is stored in the image data buffer 33 is a keyword, this pixel value is transformed to the table of other value usefulness.Therefore,, and be transformed to the table of the pixel value of 0～7 such other if pixel value map table 35a is a keyword with 0～7 such pixel value for example, just very abundant.
Here, as the pixel value of pixel value map table 35a, suppose value as shown in Figure 8.Fig. 8 is the figure of the example of remarked pixel value transform table 35a.In the table shown in this figure, as can be seen pixel value " 1 " is replaced with pixel value " 7 ", pixel value " 7 " is replaced with pixel value " 1 ".
If with reference to figure 9, then can understand the effect of this pixel value of conversion easily.Fig. 9 is the table of expression based on the pixel value before and after the conversion of view data shaping portion 35.As shown in the drawing, similar by the bit patterns that makes background (in vain) with the bit patterns of black literal lines, thus in the bit plane performance of the view data that a lot of backgrounds and literal occur, reduce with the variation of pixel value, make each bit plane as far as possible merely as target.
In the pixel value after the conversion shown in Figure 9 (transformed value), on the hypothesis attribute data, only occur under the situation of black literal lines (pixel value " 1 ") and background (pixel value " 0 "), in three bit planes that the representation attribute data are used, 2 do not occur changing fully.This is because become " 1 " by conversion and black literal lines corresponding pixel value based on the pixel value of view data shaping portion 35.
If more in general, in the view data shaping portion 35, when attribute data is decomposed into bit plane, the conversion pixel value, the possibility that makes the content of each bit plane become more simple improves.It shows as " the conversion pixel value makes to become more simple on probability " in this manual.In addition, show the possibility of certain condition of realization than the possibility height of not realizing certain condition with " on the probability " such word.That is, view data shaping portion 35 conversion pixel values, make with conversion before compare, the content of the bit plane after the conversion is more simple on probability.
Thus, because the entropy that attribute data is decomposed into each bit plane behind the bit plane reduces at least,, promptly can reduce the data volume behind the coding on probability so can improve code efficiency.Certainly this is to be prerequisite with more than half part that black literal lines and background are occupied the pixel of attribute data, but can understand under the invalid situation of this prerequisite, as long as select the high pixel value of a plurality of frequencies of occurrences, by the transform method of appropriate change pixel value, also can obtain identical effect.
Like this, encode, can improve the code efficiency of attribute data by the pixel value of replacing attribute data.
In addition,, show the map table of pixel value here, still, the transformation rule of pixel value can also be embedded into the transformation rule that medium other the display form of program shows pixel value as conditional statement with the such sheet form of pixel value map table 35a.Wherein, with the form performance of table have can the simple and reliable change of carrying out transformation rule etc. the such advantage of change.
(flow process of the processing in the view data shaping portion 35)
Figure 10 is the flow chart of an example of the flow process of the processing in the presentation video data shaping portion 35.
At first, view data shaping portion 35 obtains graph data and dwindles attribute data after portion 34 dwindles from image data buffer 33, and obtains pixel value (step S31) from this attribute data.
The 35 reference pixel value map table 35a of view data shaping portion come the pixel value (step S32) of the pixel that conversion takes out.Owing to thinking the transform method that to understand pixel value from above-mentioned explanation, so here be not described in detail.
Then, view data shaping portion 35 has judged whether to handle all pixels (step S33), if do not carry out (being not among the step S33), handles and enters into next pixel (step S34), and turn back to S31.On the other hand, if handled all pixels (among the step S33 for being), then with conversion the attribute data of pixel value output to attribute data encoding section 4, and finish a series of processing.
(structure of attribute data encoding section 4)
Figure 11 is the functional block diagram of the structure of representation attribute digital coding portion 4.As shown in the drawing, attribute data encoding section 4 has bit plane separated part 41, bit plane buffer 42 and Bit-Plane Encoding portion 43.Each inscape of attribute data encoding section 4 can be connected through data/address bus with communicating each other, and links to each other with the functional block of outside.
Bit plane separated part 41 obtains the attribute data after view data shaping portion 35 shapings of attribute data generating unit 3, and this attribute data is separated into bit plane.The attribute data that bit plane separated part 41 will be separated into bit plane stores in the bit plane buffer 42.As the method that attribute data is separated into bit plane, for example can be as shown in Figure 9, values of each row of " performance of transformed value 2 systems " is stored in the bit plane respectively.
The data that 43 pairs in Bit-Plane Encoding portion is separated into bit plane by bit plane separated part 41 and is stored in the bit plane of the attribute data in the bit plane buffer 42 are encoded.Also can use suitable prior art for this coding.Here, in order to be decomposed into bit plane, each bit plane is 2 value images, for example, JBIG (associating 2 value motion picture expert group version Joint Bi-level Image experts Group) and MMR (Modified Modified READ) etc. are enumerated coded system into its representative, can use coded system in addition.
Generally, when carrying out digital coding, because the tendency that entropy is low more, code efficiency is high more of data is arranged, regard as identical so please note the code efficiency after the entropy that will reduce usually after the conversion and the raising conversion at every turn.
(flow process of the processing in the attribute data encoding section 4)
Figure 12 is the flow chart of an example of the flow process of the processing in the representation attribute digital coding portion 4.
If obtain attribute data in the dependency data generating unit 3, then bit plane separated part 41 is separated into bit plane with this attribute data, and stores (step S41) in the bit plane buffer 42 into.
Then, 43 pairs of data that are stored in the bit plane in the bit plane buffer 42 of Bit-Plane Encoding portion are encoded, and store (step S42) in the attribute data storage part 5 into.
In addition, Bit-Plane Encoding portion 43 can not be decomposed into bit plane with attribute data, and carries out the coding as multivalue image.Be treated to the processing that is suitable for as the coding of this multivalue image under this situation in the best view data shaping portion 35.
(structure of attribute data lsb decoder 6)
Figure 13 is the functional block diagram of the structure of representation attribute data decoding part 6.As shown in the drawing, attribute data lsb decoder 6 has bit plane lsb decoder 61, bit plane buffer 62 and bit plane Synthesis Department 63.Each inscape can connect each other communicatedly through data/address bus, and links to each other with the inscape of outside.
61 pairs of the bit plane lsb decoders attribute data after by attribute data encoding section 4 codings is decoded, and stores in the bit plane buffer 62.
Comprehensive bit plane lsb decoders 61 decodings of bit plane Synthesis Department 63 also are stored in attribute data in the bit plane buffer 62, generate multivalue image.The coding/decoding method of the attribute data in the bit plane Synthesis Department 63 is that the contrary of processing in the bit plane separated part 41 handled, owing to think generally all very clear, so be not described in detail.
(flow process of the processing in the attribute data lsb decoder 6)
Figure 14 is the flow chart of an example of the flow process of the processing in the representation attribute data decoding part 6.
Obtain in the bit plane lsb decoder 61 dependency data store 5 by the attribute data behind attribute data encoding section 4 coding, the data of the bit plane of this attribute data of decoding, and store (step S51) in the bit plane buffer 62 into.
Bit plane Synthesis Department 63 obtains from bit plane buffer 62 by bit plane lsb decoder 61 decoded attribute datas, comprehensive this attribute data, and generate multivalue image (step S52).Bit plane Synthesis Department 63 stores the multivalue image that is generated in the attribute data storage part 5 into.
(structure of attribute data correction unit 7)
Figure 15 is the functional block diagram of the structure of representation attribute adjustment of data portion 7.As shown in the drawing, attribute data correction unit 7 has controller 71, view data recovery section 72, image data buffer 73 and view data enlarging section 74.Each inscape comes can connect communicatedly each other through data/address bus, and links to each other with the inscape of outside.
Each inscape of controller 71 controlled attribute adjustment of data portions 7.
Be stored in the attribute data (multivalue image) in the attribute data storage part 5 after 72 recoveries of view data recovery section are decoded by attribute data lsb decoder 6, and the attribute data after will recovering stores image data buffer 73 into.The details that the recovery in the view data recovery section 72 is handled is described in the back.
View data enlarging section 74 is amplified by view data recovery section 72 and is recovered and be stored in attribute data in the image data buffer 73.In the present embodiment,, carry out simple amplification as schematically illustrating among Figure 16 as an example of the processing in the view data enlarging section 74.Figure 16 is the figure of an example of the amplification method of the attribute data in the presentation video data enlarging section 74.Among this figure, the state after having represented to amplify attribute data state before and amplify.In the example shown in this figure, 4 pixels that constitute attribute data are enlarged into 4 times (increasing to 4) respectively, the attribute data after the amplification is made of 16 pixels altogether.In addition, in order describing, to have represented the example of the simple attribute data that constitutes by seldom pixel count among Figure 16, to have please noted that the pixel count of common formation attribute data is more.
(flow process of the processing in the attribute data correction unit 7)
Figure 17 is the flow chart of an example of the flow process of the processing in the representation attribute adjustment of data portion 7.Below in most cases, under other the situation of inscape of controller 71 control, do not put down in writing this situation especially, and be recited as the action of this inscape.
Obtain by attribute data lsb decoder 6 decoded attribute datas in the view data recovery section 72 dependency data store 5, recover this attribute data, and store (step S61) in the image data buffer 73 into.
Then, view data enlarging section 74 is obtained from image data buffer 73 by the attribute data after 72 recoveries of view data recovery section, and amplifies this attribute data (step S62).Attribute data after view data enlarging section 74 will be amplified outputs to view data efferent 13.
(recovery in the view data recovery section 72 is handled)
View data recovery section 72 has pixel value map table 72a.This pixel value map table 72a is the inverse transformation of the pixel value conversion that provides among the pixel value map table 35a that has of pixel data shaping portion 35, or has the table that the conversion of the effect approaching with it is used.
View data recovery section 72 is to decode as passing through attribute data lsb decoder 6, and the suitable value of the pixel value of the attribute data of storage is a keyword in attribute data storage part 5, and is transformed to other pixel value.Therefore,,, be transformed to the table of 0～7 such another pixel value here if be keyword with 0～7 such pixel value, just very abundant.Example with the content of the inverse transformation corresponding pixel value map table 72a of the conversion that is provided by pixel value map table 35a (with reference to figure 8) has been be provided Figure 18.
Under this situation, be simple replacement owing to used the conversion of the pixel value of pixel value map table 35a, so in the conversion of having used pixel value map table 72a, used the inverse transformation of the conversion of pixel value map table 35a.Utilize this situation, make pixel value map table 35a and pixel value map table 72a shared also no problem.Certainly, by being this situation, thinking and to bring reducing of hardware size.
(flow process of the processing in the view data recovery section 72)
Figure 19 is the flow chart of an example of the flow process of the processing in the presentation video data recovery section 72.
At first, view data recovery section 72 is from recovering by attribute data recovery section 6 and being stored in the pixel value (step S71) that takes out pixel the attribute data in the attribute data storage part 5.
The pixel value (step S72) that view data recovery section 72 reference pixel value map table 72a come conversion to take out.Owing to thinking the method for from above-mentioned explanation, having understood conversion, so be not described in detail here.
Then, view data recovery section 72 has judged whether to handle all pixels (step S73), if do not handle all pixels (being not among the step S73), handles and enters into next pixel (step S74), turns back to S71.On the other hand, if handled all pixels (among the step S73 for being), the attribute data after the recovery is stored in the image data buffer 73, and finish a series of processing.
(Change Example)
The minification that makes minification that view data dwindles the view data in the portion 8 and view data dwindle the attribute data in the portion 34 is consistent, make and on longitudinal and transverse direction, dwindle 1/2 respectively, but if the processing of image-data output device 15 is not had problems, the minification that does not then need to make view data dwindle portion 8 is consistent with the minification that view data is dwindled portion 34.
In addition, view data is dwindled in the portion 34, if " selecting the highest pixel of priority " be not even such action minification is 1/2, and is the form of 1/N, then can understand to be same action.That is, the view data minification of dwindling the attribute data in the portion 34 is not limited to 1/2.
Owing to think and the action of other inscapes that image processing apparatus 1 for present embodiment has before explanation and known technology to be interpreted as background, here so be not described in detail.
Like this, image processing apparatus 1 according to present embodiment, data at each regional attribute of the attribute of each pixel that generate, presentation video or image from view data, be in the dwindling and encode of attribute data, by carrying out pixel value conversion based on the priority of attribute, can keep image quality in images, prevent the increase of the memory capacity that storing image data is required simultaneously.
Image processing apparatus 1 can show as follows.
That is, image processing apparatus 1 has: the view data obtaining section 2 that obtains image; Comprise the view data generating unit 32 of the attribute data of property value with generation, this property value represents which in the attribute that is predetermined be a plurality of pixels of the image that composing images data obtaining section 2 obtains belong to respectively, above-mentioned property value shows as binary numerical value respectively, and this numerical value is decomposed into a plurality of bit planes, also has view data shaping portion (attribute transformation unit) 35, the above-mentioned property value of conversion makes the entropy of above-mentioned bit plane reduce, or the possibility that described at least entropy reduces improves.In addition, such performance also can show as " raising code efficiency " " to reduce entropy ".
Image processing apparatus 1 preferably have view data dwindle portion 34 and view data shaping portion 35 both, but also can have one of them.For example, do not have view data at image processing apparatus 1 and dwindle portion 34, and have under the situation of view data shaping portion 35, the be improved effect of code efficiency of attribute data of bottom line can be arranged.
In addition, attribute data gives that implication itself can understand in as described above, to not influence of operating principle of the present invention, so to have explanation here outside the attribute data of content to be suitable for the present invention also no problem.As another example of attribute data, be directly to specify in the structure in the kind of above-mentioned dither matrix rather than zone also applicable the present invention in the pixel of for example attribute data.As the different examples of attribute data, also can consider not to be pixel value, and the value conduct that makes attribute data have the kind (direction of change in concentration etc.) of representational level (gradation) is used for the information of composing images.Certainly, the content of this attribute data needs and can effectively use in view data efferent 13.
[execution mode 2]
If it is another embodiment of the present invention is described, then following such according to Figure 20～Figure 27.For the parts identical, give same Reference numeral and omit this explanation with execution mode 1.
Figure 20 is the functional block diagram of the structure of presentation video processing unit 20.As shown in the drawing, image processing apparatus 20 is different with image processing apparatus 1, has attribute data generating unit 30.
The portable image processing unit 20 of present embodiment is similar with image processing apparatus 1, but the content of fixing pixel value map table 35a according to the content of attribute data and in the adaptive transformation image processing apparatus 1.
Below, because the major part of the inscape that has of image processing apparatus 20 and image processing apparatus 1 are shared or can therefrom release easily, describe so extract understanding part required for the present invention out.
Figure 21 is the functional block diagram of the structure of representation attribute data generating unit 30.As shown in the drawing, attribute data generating unit 30 is different with attribute data generating unit 3, has view data shaping portion 36.
(structure of view data shaping portion 36)
Figure 22 is the functional block diagram of the structure of presentation video data shaping portion 36.As shown in the drawing, view data shaping portion 36 has pixel value transformation component 37, pixel value map table generating unit 38 and has stored the rewritable pixel value map table buffer 39 of pixel value map table 39a.Each inscape can connect each other communicatedly through data/address bus, and links to each other with the inscape of outside.
Pixel value transformation component 37 is extracted pixel value out from dwindle the attribute data after portion 34 dwindles by view data, and the pixel value map table 39a that generates of reference pixel value map table generating unit 38 pixel value that comes conversion to take out.Because the transform method of the pixel value in this transform method and the view data shaping portion 35 is identical, so be not described in detail here.
Pixel value map table generating unit 38 generates pixel value map table 39a, and this pixel value map table 39a is stored in the pixel value map table buffer 39.In more detail, pixel value map table generating unit 38 is come the pixel value map table 39a of conversion pixel value transformation component 37 uses according to the frequency of occurrences of the attribute of the pixel value that comprises in the attribute data.Further in other words, pixel value map table generating unit 38 conversion pixel values make that the pattern of bit plane of the pixel value that the frequency of occurrences is high is similar each other.The details of this pixel value map table generating unit 38 is described in the back.
(flow process of the processing in the view data shaping portion 36)
Figure 23 is the flow chart of an example of the flow process of the processing in the presentation video data shaping portion 36.
At first, pixel value map table generating unit 38 generates pixel value map table 39a (step S81), and stores in the pixel value map table buffer 39.
Then, pixel value transformation component 37 is obtained view data and is dwindled attribute data after portion 34 dwindles from image data buffer 33, take out pixel value (step S82) from this attribute data, and the pixel value map table 39a that generates of reference pixel value map table generating unit 38 pixel value (step S83) that comes conversion to take out.
And pixel value transformation component 37 has judged whether to handle all pixel values (step S84), if do not handle all pixel values, then handles and enters into next pixel value (step S85), and turn back to step S82.
On the other hand, if handled all pixel values, then pixel value transformation component 37 outputs to attribute data encoding section 4 (step S86) with the content of pixel value map table 39a, and finishes a series of processing.
(structure of pixel value map table generating unit 38)
Figure 24 is the functional block diagram of the structure of remarked pixel value transform table generating unit 38.As shown in the drawing, pixel value map table generating unit 38 has the order pixel value map table buffer 56 of histogram generating unit 51, histogram buffer 52, sequence list generating unit 53, sequence list buffer 54, table transform portion 55, storage order pixel value map table 56a.Each inscape can come can connect communicatedly each other through data/address bus, and links to each other with the inscape of outside.
Histogram generating unit 51 is from dwindling the pixel value that portion 34 dwindles and be stored in the attribute data the image data buffer 33 by view data, generate the histogram of the output frequency of remarked pixel value (property value), and this histogram is stored in the histogram buffer 52.That is, histogram generating unit 51 is dwindled in view data and is made the histogram that expression comprises the pixel of several which kind of attribute respectively on the attribute data after portion 34 dwindles.
The histogram that sequence list generating unit 53 generates from histogram generating unit 51 generates the sequence list of order of the frequency of occurrences of remarked pixel value, and this sequence list is stored in the sequence list buffer 54.So-called said sequence table is meant the table of the order of the frequency of occurrences that can take out each pixel value of representing by the histogram of storage in the histogram buffer 52.In other words, so-called sequence list is meant the table of order of the frequency of occurrences of the pixel value that comprises in the representation attribute data.
Table transform portion 55 is according to the order of the frequency of occurrences of each pixel value of representing of sequence list of storage in the sequence list buffer 54 and the transform method based on the pixel value of order represented by order pixel value map table 56a, generate pixel value map table 39a, and the pixel value map table 39a that is generated is stored in the pixel value map table buffer 39.
In other words, table transform portion 55 generates pixel value map table 39a, makes the order of each pixel value corresponding to the frequency of occurrences of each pixel value.
Figure 25 (a) schematically shows a histogrammic example that generates by histogram generating unit 51, and Figure 25 (b) schematically shows an example (in fact implication is the content that does not have entry name in the data of storing and be positioned at left column in histogram buffer 52 and sequence list buffer 54) of the sequence list that generates by sequence list generating unit 53.
In example shown in Figure 25, the pixel value of attribute data and the correspondence of attribute are used content shown in Figure 31, the pixel of expression black literal lines (pixel value " 7 ") is extremely few, follow the highest background (pixel value " 0 ") of the frequency of occurrences, what the frequency of occurrences was high is site (pixel value " 3 ") and colored text lines (pixel value " 4 ").
Therefore, under this situation, as carrying out in the enforcement mode 1, even the replacement pixel value, the binary system performance of the pixel value of feasible expression black literal lines shows near (it is simple that each bit plane becomes) with the binary system of the pixel value of expression background, also is difficult to code efficiency is uprised.
Otherwise, as the replacement of implementing the pixel value that carries out in the mode 1 with the frequency of occurrences in the attribute data of the pixel of the pixel of expression background (pixel value " 0 ") and expression black literal lines (pixel value " 7 ") any one all high be prerequisite, still this prerequisite is false in the described situation of present embodiment.
In this case, by content according to attribute data, more particularly, come the transform method of adaptive determining attribute data according to the probability distribution (histogram) of the pixel value of attribute data, the content of each bit plane after the Bit Plane Decomposition of attribute data becomes merely, and entropy reduces on probability at least.Therefore, code efficiency improves, and can be characteristics of the present invention in encoded attributes data under the high efficiency.
Order pixel value map table 56a be pixel value the frequency of occurrences order and be in conversion after the correspondence table of pixel value of attribute data.Figure 26 schematically shows the example of this order pixel value map table 56a.Because existing the helping of binary form understood, thus introduce, but it is not the part of table.Also identical to left column with entry name.
In example shown in Figure 26, making the order former state is pixel value, but in addition the transform method of pixel value is also considered.As mentioned above, be similar each other binary system performance by making the high pixel value of the frequency of occurrences after Bit Plane Decomposition, each bit plane after the Bit Plane Decomposition becomes simple, therefore, the entropy of each bit plane reduces, and code efficiency improves, and promptly the data volume behind the coding reduces most important.If can realize the transform method of this purpose, then the transform method of the pixel value map table 39a in the table transform portion 55 can be so like this.
(flow process of the processing in the pixel value map table generating unit 38)
Figure 27 is the flow chart of an example of the flow process of the processing in the remarked pixel value transform table generating unit 38.
At first, histogram generating unit 51 obtains from image data buffer 33 by view data dwindles attribute data after portion 34 dwindles, from this attribute data, generate the histogram of the frequency of occurrences of the pixel value that comprises in this attribute data of expression, and store (step S91) in the histogram buffer 52 into.
In addition, sequence list generating unit 53 obtains the histogram that histogram generating unit 51 generates from histogram buffer 52, genesis sequence table from this histogram, and store (step S92) in the sequence list buffer 54 into.
Table transform portion 55 is according to the order and the order pixel value map table 55a of the frequency of occurrences of each pixel value that takes out from sequence list buffer 54, generate pixel value map table 39a, and the pixel value map table 39a that is generated is stored in the pixel value map table buffer 39 (step S93).
Pixel value transformation component 37 is in step S86 (Figure 23), the content of pixel value map table 39a is outputed in the attribute data encoding section 4, but this is because image processing apparatus 20 effective actions in order to make present embodiment, except that before illustrated, need on the image processing apparatus 1 of execution mode 1, give this change of inverse transformation of the conversion carried out in the view data recovery section 72 (conversion of representing by pixel value map table 72a).But, because for the content with pixel value map table 39a passes to view data recovery section 72, and in view data recovery section 72, carry out this inverse transformation, from explanation before, find out concerning those skilled in the art and be easy to thereby change image processing apparatus 1, so omit above details.
As mentioned above, image processing apparatus 20 is to comprise image obtaining section (obtaining the unit) 2 that obtains image and the image processing apparatus that generates the view data generating unit (attribute data generation unit) 32 of the attribute data that comprises property value, wherein this property value represents which in the attribute that is predetermined be a plurality of pixels of the image that composing images data obtaining section 2 obtains belong to respectively, above-mentioned property value shows as binary numerical value respectively, and this numerical value is decomposed into a plurality of bit planes, this image processing apparatus comprises: each property value that comprises in the attribute data that generates at view data generating unit 32, calculate the histogram generating unit (frequency of occurrences arithmetic element) 51 of the frequency of occurrences of each property value; With the frequency of occurrences according to the property value of calculating by histogram generating unit 51, the above-mentioned property value of conversion makes the entropy of above-mentioned each bit plane reduce, or the possibility that reduces of the described at least entropy table transform portion (property value converter unit) 55 of improving.
In other words, table transform portion this property value of 55 conversion makes that the pattern of bit plane of the property value that the frequency of occurrences in the property value that attribute data comprised is high is similar each other.
In addition, image processing apparatus 20 further comprises the frequency of occurrences of calculating according to histogram generating unit 51, come the sequence list generating unit (order calculate unit) 53 of order of the frequency of occurrences of computation attribute value, the order that table transform portion 55 calculates according to sequence list generating unit 53 changes above-mentioned property value.
Table transform portion 55 uses the numerical value of the order that order of representation table generating unit 53 is calculated as property value.
(Change Example)
The present invention is not limited to each above-mentioned execution mode, can in the scope shown in the claim, carry out various changes, but the execution mode that obtains for disclosed respectively technological means in the different execution mode of appropriate combination is also contained in the technical scope of the present invention.
Each piece of above-mentioned image processing apparatus 1 and image processing apparatus 20, especially attribute data generating unit 3 and 30 can constitute by hardware, also can be following such, and use CPU to realize by software.
Promptly, image processing apparatus 1 and image processing apparatus 20 comprise the CPU (CPU of the order of carrying out the control program of realizing each function, central processing unit), stored the ROM (read-only memory of said procedure, read only memory), launch the storage device (recording medium) etc. of RAM (random access storage device, random access memory), storage said procedure and the various memory of data etc. of said procedure.And, can offer above-mentioned image processing apparatus 1 and image processing apparatus 20 by the recording medium that computer has write down as the procedure code (execute form program, intermediate code program, software program) of the control program (image processing program) of the image processing apparatus of the software of realizing above-mentioned functions and image processing apparatus 20 with reading, also carried out by the procedure code that this computer (or CPU or MPU) writes down on the read-out recording medium, thereby can be realized purpose of the present invention.
As aforementioned recording medium, can use semiconductor memory series of the card series of dish series, IC-card (comprising storage card)/light-card etc. of CD of band series, the disk that comprises floppy disk (registered trade mark)/hard disk etc. and CD-ROM/MO/MD/DVD/CD-R etc. of for example tape and tape etc. or mask rom/EPROM/EEPROM/ flash rom etc. etc.
Image processing apparatus 1 and image processing apparatus 20 constitute joinably with communication network, can supply with the said procedure sign indicating number through communication network.As this communication network, be not particularly limited, but can use internet, Intranet, wide area network, LAN, ISDN, VAN, CATV communication network, Virtual Private Network (virtual private network), telephone wire road network, mobile radio communication, satellite communication network etc.In addition, do not limit especially as the transfer medium that constitutes communication network, for example, not only in IEEE1394, USB, power line transmission, wired TV net, telephone wire, adsl line etc. wired, and in the such infrared ray of IrDA and remote control, bluetooth (registered trade mark), 802.11 wireless, HDR, mobile telephone network, satellite circuit, ground digital net etc. wireless, also can use.The present invention also can transmit the said procedure sign indicating number and specific the realizing of quilt by electronics under the form of the computer data signal that embeds on the carrier wave.
In addition, in the above-mentioned explanation, the configuration example of entropy of the bit plane of the pixel value that reduces the representation attribute data on probability has been described, but also can realize following formation: the entropy by the bit plane after the pixel value of measuring the conversion attribute data is also regulated the transform method of pixel value, thereby guarantee to reduce the entropy after the conversion, or with conversion before compare at least and do not increase.Can understand that this structure can be realized by the change of giving seldom on said structure.
Concrete, for example can further be provided with: measuring unit, its measure expression by view data shaping portion 35 or 36 conversion of view data shaping portion the entropy of bit plane of pixel value of the attribute data before and after this pixel value; With the change unit, it compares and measures the above-mentioned entropy of the conversion front and back of the pixel value of measuring the unit, under the big situation of the above-mentioned entropy of above-mentioned entropy after conversion before, change the transform method of the pixel value of the attribute data in view data shaping portion 35 or the view data shaping portion 36 than conversion.
In this case, can really admit a fault on probability, but the conversion by pixel value reduces entropy or entropy is not increased.
In addition, also can consider by the data volume behind the measurement coding, rather than measure entropy, guarantee to improve before code efficiency is than conversion, or the method that does not reduce at least.In this case, confirmed it is not on probability, but, code efficiency has been improved, or do not reduce by conversion.Certainly, the method for measurement of the data volume behind the coding in this case need be complementary with the coding method of use in the Bit-Plane Encoding portion 43.
The present invention can show as follows.
That is, image processing apparatus of the present invention is characterised in that, comprising: view data input unit, its input image data; Image data memory cell, it stores described view data; The attribute data generation unit, it generates the attribute data of the attribute of each pixel of representing composing images according to described view data; The attribute data memory cell, it stores described attribute data; Image data-outputting unit, it is with reference to the described attribute data of storing in the described attribute data memory cell, and output is stored in the described view data in the described image data memory cell; Described attribute data generation unit dwindles described attribute data according to the priority that offers each pixel value.
Preferred described attribute data generation unit keeps each priority that provides of described each pixel value with form.
Image processing apparatus of the present invention is characterised in that, comprising: view data input unit, its input image data; Image data memory cell, it stores described view data; The attribute data generation unit, it generates the attribute data of the attribute of each pixel of representing composing images according to described view data; The attribute data coding unit, its described attribute data of encoding; The attribute data memory cell, it stores the attribute data behind the described coding; Image data-outputting unit, the described attribute data that it is stored with reference to described attribute data memory cell, output is stored in the described view data in the described image data memory cell; The pixel value of described attribute data generation unit conversion attribute data, make the entropy of each bit plane when described attribute data is decomposed into bit plane on probability, reduce at least, or the pixel value of described attribute data generation unit conversion attribute data, make the code efficiency that is decomposed under the situation of bit plane at described attribute data on probability, improve at least.
Preferred described attribute data generation unit is according to the content of attribute data and the transform method of the pixel value of the described attribute data of adaptive determining.
Preferred described attribute data generation unit is according to the probability distribution of the pixel value of attribute data, the transform method of the pixel value of the described attribute data of adaptive determining.
Image processing method of the present invention is characterised in that, comprising: the step of input image data; Store the step of described view data; According to described view data, the step of the attribute data of the attribute of each pixel of this image of generation expression formation; Store the step of described attribute data; With reference to described attribute data, export the step of described view data, in the step that generates described attribute data, dwindle described attribute data according to the priority of providing for each pixel value.
Image processing method of the present invention is characterised in that, comprising: the step of input image data; Store the step of described view data; According to described view data, the step of the attribute data of the attribute of each pixel of this image of generation expression formation; The encode step of described attribute data; Store the step of the attribute data behind the described coding; With reference to described attribute data, export the step of described view data, in the step that generates described attribute data, the pixel value of conversion attribute data, make the entropy of each bit plane when described attribute data is decomposed into bit plane on probability, reduce at least, perhaps in the step that generates described attribute data, the pixel value of conversion attribute data makes that the code efficiency under the situation that described attribute data is decomposed into bit plane improves at least on probability.
Storage medium stores of the present invention be used to make computer to carry out the program of following steps, that is: the step of input image data; Store the step of described view data; According to described view data, the step of the attribute data of the attribute of each pixel of generation expression composing images; Store the step of described attribute data; With the described attribute data of reference, export the step of described view data, it is characterized in that, in the step that generates described attribute data, dwindle described attribute data according to the priority that each pixel value is provided.
Medium of the present invention has been stored and has been used to make computer to carry out the program of following steps, that is: the step of input image data; Store the step of described view data; According to described view data, the step of the attribute data of the attribute of each pixel of generation expression composing images; The encode step of described attribute data; Store the step of the attribute data behind the described coding; With reference to described attribute data, export the step of described view data, it is characterized in that, in the step that generates described attribute data, the pixel value of conversion attribute data makes the entropy of each bit plane when described attribute data is decomposed into bit plane reduce on probability at least, or in the step that generates described attribute data, the pixel value of conversion attribute data makes that the code efficiency under the situation that described attribute data is decomposed into bit plane improves at least on probability.
As mentioned above, among the present invention above-mentioned attribute data dwindle the unit the highest property value of priority in the property value of handy these a plurality of pixels replace the property value of a plurality of pixels that predetermined zone comprised in the above-mentioned image.
According to said structure, attribute data dwindles the unit and replaces the property value of a plurality of pixels that predetermined zone comprised in the image of obtaining the pictorial data representation of obtaining the unit with the highest property value of property value medium priority of a plurality of pixels that comprise in the described predetermined zone.
So-called above-mentioned predetermined zone is meant by cutting apart at least one cut zone in a plurality of cut zone that the image of for example obtaining the pictorial data representation of obtaining the unit produces.
Therefore, can with a property value corresponding with the highest pixel of priority replace with presumptive area in the corresponding a plurality of property values of a plurality of pixels that comprise, can dwindle attribute data when keeping the highest property value of priority reliable.
In addition, above-mentioned priority list is shown priority list, it is the priority list of the above-mentioned attribute of expression and the corresponding relation of the value of the priority of this attribute of expression, and above-mentioned attribute data dwindles the unit preferably by with reference to above-mentioned priority list, judges the priority of above-mentioned attribute.
According to said structure, attribute data dwindles the unit and judges the priority that the attribute of pixel is given by the reference priority list.
Therefore, can easily carry out the change of priority by changing the priority that priority list changes the attribute of pixel.
In addition, image processing apparatus of the present invention is characterised in that, comprising: obtain the unit, it obtains view data; The attribute data generation unit, its generation comprises the attribute data of property value, and this property value is represented to constitute the above-mentioned a plurality of pixels that obtain the image of the pictorial data representation of obtaining the unit and is belonged in the attribute that is predetermined which respectively; Priority decision unit, its according to the content of the interior perhaps above-mentioned attribute data of above-mentioned view data at least one of them, decide the priority that above-mentioned each property value is given; Dwindle the unit with attribute data, it is transformed to a property value according to the priority by above-mentioned priority decision unit decision by at least two property values that comprise in the attribute data that above-mentioned attribute data generation unit is generated, and dwindles this attribute data.
According to said structure, the attribute data generation unit generates the attribute data comprise property value, and this property value represents which attribute is a plurality of pixels that formation obtains the image of the pictorial data representation of obtaining the unit belong to respectively.Priority decision unit according to the content of the interior perhaps attribute data of view data at least one of them, decide the priority that each property value is given.And attribute data dwindles the priority of unit according to priority decision unit decision, is transformed to a property value by at least two property values that comprise in the attribute data that the attribute data generation unit is generated and dwindles attribute data.
Therefore, can in the high property value of reliable reservation priority, dwindle attribute data.
Image processing method of the present invention is to comprise the image processing method of obtaining in the image processing apparatus that unit, attribute data generation unit and attribute data dwindle the unit, it is characterized in that comprise: obtain step, view data is obtained in the above-mentioned unit of obtaining; Attribute data generates step, and above-mentioned attribute data generation unit generates the attribute data that comprises property value, and property value is represented to constitute the above-mentioned a plurality of pixels that obtain the image of the pictorial data representation of obtaining in the step and belonged in the attribute that is predetermined which respectively; The attribute data reduction process, according to the priority that above-mentioned each property value is given in advance, be transformed to a property value by above-mentioned attribute data being generated at least two property values that comprise in the attribute data that generates in the step, dwindle described attribute data thereby above-mentioned attribute data dwindles the unit.
A kind of image processing program that makes above-mentioned image processing apparatus action is even computer is also contained in the technical scope of the present invention as above-mentioned each unit image processing program that works and the recording medium that has write down the embodied on computer readable of this image processing program.
The present invention is because reliable important property value in the property value that comprises in the reserved property data, and the while can effectively be dwindled this attribute data, so can be suitable for well as the image processing apparatus of use attribute data.
Concrete execution mode of having done in the detailed explanation of invention or embodiment are the content for technology contents of the present invention is understood all the time, do not answer narrow definition for only limiting to this object lesson, spirit of the present invention and below in the scope of claims of putting down in writing, can carry out various changes and implemented.

Claims (5)

1, a kind of image processing apparatus comprises:
Obtain the unit, it obtains view data;
The attribute data generation unit, its generation comprises the attribute data of property value, and this property value is represented to constitute the described a plurality of pixels that obtain the image of the pictorial data representation of obtaining the unit and is belonged in the attribute that is predetermined which respectively; With
Attribute data dwindles the unit, and it is transformed to a property value by at least two property values that comprise in the attribute data that described attribute data generation unit is generated, thereby dwindles this attribute data according to the priority that described each property value is given.
2, image processing apparatus according to claim 1 is characterized in that,
Described attribute data dwindles the unit and replaces the property value of these a plurality of pixels with the highest property value of property value medium priority of a plurality of pixels that predetermined zone comprised in the described image.
3, image processing apparatus according to claim 1 is characterized in that,
Described priority is represented as: represents described property value and represents the priority list of the corresponding relation between the value of priority of this property value,
Described attribute data dwindles the unit by judge the priority of described property value with reference to described priority list.
4, image processing apparatus according to claim 1 is characterized in that,
Also have the priority generation unit, it is at least a according to the content of the interior perhaps described attribute data of described view data, decides the priority that described each property value is given;
Described attribute data dwindles the unit according to dwindling described attribute data by the priority of described priority generation unit decision.
5, the image processing method in a kind of image processing apparatus comprises:
Obtain step, obtain view data;
Attribute data generates step, generates the attribute data that comprises property value, and this property value is represented to constitute the described a plurality of pixels that obtain the image of the pictorial data representation of obtaining in the step and belonged in the attribute that is predetermined which respectively; With
The attribute data reduction process according to the priority that described each property value is given, is transformed to a property value by described attribute data is generated at least two property values that attribute data comprised that generate in the step, thereby dwindles described attribute data.
CNA2008100878720A 2007-03-30 2008-03-27 Image processing device, image processing method Pending CN101277366A (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
JP2007-094987 2007-03-30
JP2007094987 2007-03-30
JP2008-033685 2008-02-14

Publications (1)

Publication Number Publication Date
CN101277366A true CN101277366A (en) 2008-10-01

Family

ID=39996347

Family Applications (1)

Application Number Title Priority Date Filing Date
CNA2008100878720A Pending CN101277366A (en) 2007-03-30 2008-03-27 Image processing device, image processing method

Country Status (2)

Country Link
JP (1) JP2008278462A (en)
CN (1) CN101277366A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107005657A (en) * 2016-10-13 2017-08-01 深圳市大疆创新科技有限公司 Method, device, chip and the camera of processing data
CN110140164A (en) * 2016-12-27 2019-08-16 硅工厂股份有限公司 The sensing circuit of source electrode driver and the display equipment for using it

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107005657A (en) * 2016-10-13 2017-08-01 深圳市大疆创新科技有限公司 Method, device, chip and the camera of processing data
US10855917B2 (en) 2016-10-13 2020-12-01 SZ DJI Technology Co., Ltd. Data processing method and device, chip, and camera
CN110140164A (en) * 2016-12-27 2019-08-16 硅工厂股份有限公司 The sensing circuit of source electrode driver and the display equipment for using it

Also Published As

Publication number Publication date
JP2008278462A (en) 2008-11-13

Similar Documents

Publication Publication Date Title
Baharav et al. Watermarking of dither halftoned images
CN102413265B (en) Image processing apparatus and control method thereof
US20080175476A1 (en) Apparatus and method of segmenting an image and/or receiving a signal representing the segmented image in an image coding and/or decoding system
US8503036B2 (en) System and method of improving image quality in digital image scanning and printing by reducing noise in output image data
CN101800827B (en) Apparatus and method
JP2007336226A (en) Information processor, control method, and computer program
CN104054047A (en) Image processing device, image forming device, program and storage medium
US7426054B1 (en) Image processing apparatus, image reproduction apparatus, system, method and storage medium for image processing and image reproduction
JP4364809B2 (en) Image processing apparatus, image processing method, program, and recording medium
JP5893379B2 (en) Image compression apparatus, image compression method, and computer program
US8107743B2 (en) Image processing device, image processing method, and storage medium
CN114332479A (en) Training method of target detection model and related device
CN101277366A (en) Image processing device, image processing method
JPH11168632A (en) Binary expression processing method for dither image, method for uncompressing dither image expressed in compression binary representation and compression and uncompression system for dither image
JP4371026B2 (en) Printing system and printer
US7259891B2 (en) Image compression method, decompression method thereof and program therefor
Saraswat et al. A study on size optimization of scanned textual documents
JPH04356873A (en) Adaptive encoding system for color document image
JP3675429B2 (en) Adaptive predictive encoding and decoding method and apparatus thereof, and recording medium recording adaptive predictive encoding and decoding program
JP3404380B2 (en) Adaptive predictive encoding / decoding method and apparatus, and recording medium storing adaptive predictive encoding / decoding program
JP2004088251A (en) Image processing apparatus and method, program for executing the method, and recording medium
JP4757172B2 (en) Image coding apparatus and control method thereof
JP4719924B2 (en) Image processing apparatus and image processing method
JP7185451B2 (en) Image processing device, image processing method, and program
JP2014135574A (en) Image processing device, encoding method, and decoding method

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C12 Rejection of a patent application after its publication
RJ01 Rejection of invention patent application after publication

Open date: 20081001